Efficient model for prediction the adsorption of organic compounds by single-walled carbon nanotube
کد مقاله : 1120-INZC6
ابوذر خواجه *
گروه مهندسی شیمی، دانشگاه صنعتی بیرجند، بیرجند، ایران
چکیده مقاله:
Carbon nanotubes have been identified as potential adsorbents due to their unique and outstanding structure and large surface area. In this work, modified particle swarm optimization-multiple linear regression (MPSO-MLR) [1, 2] was applied to the quantitative structure-property relationship (QSPR) modeling of adsorption of 37 diverse chemical compounds on single-walled carbon nanotube (SWCNT). For developing QSPR model, the data set was collected from [3], and after the calculation a variety of descriptors from the optimized molecular structures, a subset of relevant descriptors was selected and the optimal linear model was developed as follows:

LogK = 0.6134 ×ATS3v + 0.7039 ×EEig04x - 2.0841 ×C-002+ 7.2007 ×BELe4 - 6.9193 ×BELv4+ 1.1557
N=37; R2= 0.925; R2adj =0.913; Q2LOO=0.902.

The interpretation of molecular descriptors used in linear model indicates distribution of atomic volumes and electronegativity along topological molecular structure of compounds increases the adsorption of compounds on SWCNT. Moreover existence the CH2R2 atom-centred fragments causes a decreases in the adsorption.
کلیدواژه ها:
QSPR; Single-walled carbon nanotube; Adsorption; MPSO-MLR
وضعیت : مقاله پذیرفته شده است